I call this one the tester than knew too much.

Note: The server the tester is running on has a hard drive that is breaking down...

Tester: Remember the error I talked to you about yesterday?

Me: Yeah, what about it?

Tester: Well the server hasn't recovered yet and I haven't restarted anything...

Me: Well the application itself hasn't crashed so our monitoring application doesn't seem to notice that the service is in a bad state. The error seems only to have brought down certain threads within the application.

Tester: No, I think there is a different issue here and has nothing to do with that error, the application is still doing things.
*tails the log*
Tester: See?

Me: As I said some things are still running and are unaffected by the error.

Tester: NO! It has to be caused by the other error I had a week ago where my file got corrupted. As we said I removed the file, restarted it and it worked again, but had the same problem a day later...

Note: The problem is not related, this time the application is running out of file descriptors

Me: Well... If the problem is the same it would have complained about the file descriptors then aswell, not an I/O error.

Tester: Nope, I think you are wrong!

Me: ¯\_(ツ)_/¯

FML